Angewandte
Chemie
DOI: 10.1002/anie.201001938
“Heavy” Hydrogen Cyanide
Bonding in the Heavy Analogue of Hydrogen Cyanide: The Curious
Case of Bridged HPSi**
Valerio Lattanzi, Sven Thorwirth, DeWayne T. Halfen, Leonie Anna M!ck, Lucy M. Ziurys,
Patrick Thaddeus, J!rgen Gauss,* and Michael C. McCarthy*
The bonding of first- and second-row elements differ dramatically. The simplest unsaturated silicon hydrides Si2H2 and
Si2H4 exhibit quite unusual geometries[1] compared to the
analogous hydrocarbon molecules. For example, the most
stable form of Si2H2 is nonplanar with C2v symmetry and two
bridging H atoms, in sharp contrast to linear acetylene, HC!
CH. Phosphorus and nitrogen share many of the same
bonding characteristics, but P prefers single over multiple
bonds. For these reasons, it may be difficult to predict the
most stable isomeric arrangement, even for a small molecule
with a single P or Si atom and especially when it contains both.
Silicon–phosphorus bonds are important in materials
science[2] and organometallic chemistry.[3] Geometries for
several-hundred larger molecules containing a phosphorus–
silicon single bond have been determined by X-ray crystallography,[4] but the uncertainties are typically 0.05 !. With
respect to phosphorus–silicon multiple bonds, it appears that
sterically stabilized phosphasilenes[5] R2Si = PR’ are the only
[*] Dr. V. Lattanzi, Prof. P. Thaddeus, Dr. M. C. McCarthy
Harvard-Smithsonian Center for Astrophysics
60 Garden Street, Cambridge, MA 02138 (USA)
and
School of Engineering and Applied Sciences, Harvard University
Cambridge, MA 02138 (USA)
Fax: (+ 1) 617-495-7013
E-mail: [email protected]
Dr. S. Thorwirth
I. Physikalisches Institut, Universit!t zu K"ln
50937 K"ln (Germany)
and
Max Planck Institut f#r Radioastronomie
53121 Bonn (Germany)
Dr. D. T. Halfen, Prof. L. M. Ziurys
Departments of Chemistry and Astronomy and Steward Observatory, University of Arizona
Tucson, AZ 85721 (USA)
Dipl.-Chem. L. A. M#ck, Prof. Dr. J. Gauss
Institut f#r Physikalische Chemie, Universit!t Mainz
55099 Mainz (Germany)
Fax: (+ 49) 6131-39-23895
E-mail: [email protected]
[**] Financial support for the work in Cambridge and Tucson is provided
by the National Science Foundation (NSF) Center for Chemical
Innovation (CCI) grant CHE 08-47919. D.T.H. is supported by a NSF
Astronomy and Astrophysics Post-doctoral Fellowship (AST 0602282). S.T. and J.G. gratefully acknowledge funding by the
Deutsche Forschungsgemeinschaft (DFG) through grants TH1301/
3-1 and GA370/5-1. L.A.M. is supported by a fellowship by the
Graduate School of Excellence “MAINZ”.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201001938.
1
! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
!
!
Angew. Chem. Int. Ed. 2010, 49, 1 – 5
compounds that have been isolated to date. The precise
structures have been determined for only two of these species,
using single-crystal X-ray crystallography.[6] Our knowledge
of the Si"P bond is therefore quite inadequate.
Gas-phase investigations of phosphorus–silicon compounds have been hampered by their high reactivity;
consequently the structures of only a few species with Si"P
bond have been accurately characterized,[7] that is, with
accuracies of at least 0.05 !. Nevertheless, it should be noted
that small molecules containing both elements are also of
astronomical interest, because silicon- and phosphorus-bearing molecules[8j have been detected in the interstellar space by
radio astronomy.
HPSi is perhaps the simplest unsaturated compound to
have a chemical bond between silicon and phosphorus.
Quantum-chemical calculations and experimental investigations[9] have shown that linear structures are by far the most
stable arrangements for HCN, HNC, and the heavier analogues HNSi and HCP. Until the present investigation, no
experimental data were available for HPSi, but ab initio
calculations[10–12] concluded that a bridged structure with a Si"
P double bond (hereafter denoted HPSi) is more stable than
linear HSiP. These calculations[12] also predict an energy
difference, of the order of 10 kcal mol"1, between these forms;
a transition state lying 13 kcal mol"1 above HSiP connects the
two isomers. Although hydrogen bonding occurs in silicon
hydrides, no phosphorus-bearing molecules with this type of
bonding are known.
To determine the existence, geometry, and bonding of
HPSi, a study has been undertaken to measure its rotational
spectrum. Rotational spectroscopy is an ideal technique to
study HPSi and other polar molecules because rotational
frequencies are directly related to the moments of inertia
along the three principal axes of the molecule. With sufficient
isotopic substitutions, it is then possible to determine highly
accurate molecular structures, that is, with bond lengths to an
uncertainty on the order of 0.010 !. When the experimental
rotational constants are corrected for zero-point vibrational
effects, it is possible to derive bond lengths to even higher
accuracy, on the order of # 0.001 ! or better. The experimental investigations presented herein were guided by highlevel coupled-cluster calculations at the same level of theory
as that reported recently for silanethione, H2SiS.[13] Details
concerning the approach used to treat electron-correlation
and basis-set effects in quantum-chemical calculations in a
quantitative manner are given below and can also be found in
Ref. [14].
HPSi was produced in the gas phase through a discharge
of silane and phosphine in neon and then detected at high-
These are not the final page numbers!
Communications
spectral resolution by means of Fourier transform microwave
and millimeter wave absorption spectroscopy. The existence
of HPSi was established by the observation of its lowest
energy rotational transition near a frequency of 16 GHz (see
Figure 1). The spectrum shows closely spaced hyperfine
Table 1: Rotational constants of HPSi and DPSi (in MHz).[a]
Constant
A0
B0
C0
HPSi
DPSi
Exp.
Calcd[b]
Exp.
Calcd[b]
297 187(20)
8168.95742(81)
7936.6581(11)
296 559.5
8170.0
7939.0
152 598(63)
8169.9398(16)
7737.4098(16)
152 442.0
8171.5
7740.0
[a] 1s uncertainties (in parentheses) are in units of the last significant
digit. [b] See text.
Figure 1. The lowest energy (10,1–00,0) rotational transition of HPSi
showing resolved hyperfine structure from the interaction of the 31P
nuclear spin I = 1=2 with the small magnetic field induced by rotation. A
best-fit stick spectrum in gray is shown beneath that observed, with
the calculated relative intensities. Each rotational transition has a
double-peak line shape, because of the Doppler doubling caused by
the Mach 2 molecular beam expanding collinear to the two propagating directions of the molecular emission field. The integration time
was 1 min.
2
!
!
structure, characteristic of a molecule containing 31P with
spin I = 1=2 . Because of the high signal-to-noise ratio, the
corresponding lines of the 29Si and 30Si isotopic species were
also detected nearby in frequency, despite the low fractional
abundance of these two isotopes (4.7 % for 29Si and 3.1 % for
30
Si). Additional lines, nearly harmonic in frequency to these,
were soon detected near 32 GHz. By replacing silane with
SiD4 the same two transitions were detected for DPSi, DP29Si,
and DP30Si. Because D substitution lowers the A rotational
constant by a factor of two, it was also possible to measure two
additional lines of DPSi, which arise from Ka = 1, as opposed
to Ka = 0. After the 16 to 32 GHz measurements, the rotational spectra of two isotopic species, HPSi and DPSi, were
subsequently observed at substantially higher frequencies
(287–421 GHz) to better determine their spectroscopic constants of both isotopic species.
More than 200 rotational transitions have now been
measured for HPSi, and 30 for DPSi. Spectroscopic constants
for both isotopic species were then determined from a nonlinear least-squares analysis using a standard asymmetric-top
Hamiltonian (see Supporting Information for complete data
sets). The data for both species could be reproduced with the
three rotational constants, A, B, and C, and five centrifugal
distortion constants, to a root mean square (rms) accuracy of
only 40 kHz, that is, a fraction of the spectral line width for
most of the measurements. The rotational constants, each
determined to better than 0.1 %, are summarized in Table 1
and there compared to the theoretical predictions. A more
complete account of the laboratory measurements, data
www.angewandte.org
analysis, and quantum-chemical calculations will be given
elsewhere.
Because the three rotational constants are inversely
proportional to the principal moments of the molecule, it is
possible to precisely determine the positions of all three
nuclei by a least-squares minimization of the HPSi and DPSi
data. To determine the most accurate structure, the experimental rotational constants were corrected for zero-point
vibration, a small but important correction, prior to minimization. The structure thus derived is shown in Figure 2. With
the present data set, the three bond lengths are each
determined to better than 0.006 !, the ]SiPH angle to
better than 0.58.
Figure 2. Molecular structure of bridged HPSi derived from the experimental rotational constants of HPSi and DPSi corrected for zero-point
vibrational effects (see text). Statistical uncertainties in units of the
last significant digit are in parentheses. The theoretical structure
obtained using coupled-cluster methods (see text) is given in italics
for comparison; the calculated dipole moment components along the
two principal axes shown are ma = 0.89 D and mb = 0.25 D.
From a purely geometrical perspective, the SiP bond in
HPSi appears to be best characterized as a strong Si=P double
bond; it is 0.05 ! shorter than the value (2.094 !) determined
by X-ray crystallography on phosphasilenes,[4] but is 0.05 !
longer than that expected for a triple bond (1.993 !).[10] The
P"H distance is 5 % longer than a normal P"H bond (i.e. PH3,
1.41 !).[15] The Si"H distance (1.843 !) is roughly 0.4 !
longer than in SiH4 (1.471 ![16]) and 0.2 ! longer than in
bridged Si2H2 (1.668 ![1]), but much shorter than the sum of
the van der Waals radii (3.2 !) implying significant bonding.
A natural bond orbital (NBO) analysis[17] yields two bonding
orbitals for the P"Si bond and one for P"H. However, no
normal bonding orbital is found for H"Si. In addition, two
lone-pair orbitals are identified, one at P and one at Si.
In view of these findings, HPSi does not adopt a cyclic but
a bent structure, although the H-P-Si angle of 60.5(3) degrees
! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers!
Angew. Chem. Int. Ed. 2010, 49, 1 – 5
Angewandte
Chemie
clearly indicates significant PH–Si interactions. The bonding
in HPSi can be summarized as in Figure 3 with a classical P=Si
double bond, a P"H bond, and donor–acceptor interactions
between PH and the empty p orbital at the Si atom.
ducted on circumstellar gas using the Arizona Radio Observatory 12 m telescope. The results of these observations will be
described elsewhere.
Experimental Section
Contrary to HCP and HNSi, it cannot be as easily
explained by qualitative arguments, why the HPSi arrangement is preferred over the HSiP form. Linear HPSi, which is
not a minimum on the potential energy surface, is slightly less
stable (about 1 kcal mol"1) than linear HSiP, and the preference for the HPSi arrangement is solely due to the bending.
The bent structure of HPSi can be rationalized by the fact that
P, because of a lower tendency to form hybrid orbitals[18]
favors arrangements with a lone pair over those with a triple
bond. In the present case, P avoids a triple bond and instead
adopts a bonding situation with a double bond and a lone pair
of essentially s-orbital character. We emphasize the importance of the PH–Si donor–acceptor interactions, as they are
essential for the stabilization of the non-linear structure.
It is remarkable that the most stable arrangement of H, Si,
and P is bridged when the three isovalent analogues of HPSi
(HCN, HNSi, and HCP) are linear. Calculations suggest that
HPSi may be fairly unique among SiP compounds, as no other
SiP compound with a bridged structure has been experimentally characterized to date. Similar examples have been
discussed solely on the basis of quantum-chemical calculations, namely the P2H+ ion with a cyclic structure and HPPR
with electron-withdrawing substituents R.[19] In light of the
present findings, it is clearly desirable to characterize
experimentally the structures of other species of greater
complexity, such as HPSiH2.[10, 12] As the recent experimental
detection of a fairly stable, new bridged isomer of Si2H4
demonstrates,[1c] isomers with bridging H atoms may exist
for other simple SiP compounds as well. Molecules such as
H2PSiH and HPSiH2, may be detectable with the present
technique because HPSi is efficiently produced using PH3 as a
source of P and silane as a source of Si. If so, it should be
possible to determine, at high spectral resolution, the
structure and bonding properties of a number of fundamental
SiP bearing molecules.
Because HCN, HCP, and other small silicon- and phosphorus-bearing compounds (i.e. CCP, PO) have been detected
in circumstellar shells of evolved carbon-rich stars[8] and other
astronomical sources, HPSi is also a good candidate for radio
astronomical detection. Searches for HPSi have been conAngew. Chem. Int. Ed. 2010, 49, 1 – 5
Received: April 1, 2010
Revised: May 8, 2010
Published online: && &&, 2010
.
Keywords: multiple bonds · phosphorous ·
quantum-chemical calculations · rotational spectroscopy ·
silicon
[1] a) M. Bogey, H. Bolvin, C. Demuynck, J. L. Destombes, Phys.
Rev. Lett. 1991, 66, 413 – 416; b) M. Cordonnier, H. Bolvin, C.
Demuynck, J.-L. Destombes, J. Chem. Phys. 1992, 97, 7984 –
7989; c) L. Sari, M. C. McCarthy, H. F. Schaefer III, P. Thaddeus,
J. Am. Chem. Soc. 2003, 125, 11409 – 11417.
[2] J. M. Jasinski, B. S. Myerson, B. A. Scott, Annu. Rev. Phys.
Chem. 1987, 38, 109 – 140.
[3] G. Fritz, P. Scheer, Chem. Rev. 2000, 100, 3341 – 3401.
[4] F. H. Allen, J. E. Davies, J. J. Galloy, O. Johnson, O. Kennard,
C. F. Macrae, E. M. Mitchell, G. F. Mitchell, J. M. Smith, D. G.
Watson, J. Chem. Inf. Comput. Sci. 1991, 31, 187 – 204.
[5] a) C. N. Smith, F. Bickelhaupt, Tetrahedron Lett. 1984, 25, 3011 –
3014; b) C. N. Smit, F. Bickelhaupt, Organometallics 1987, 6,
1156 – 1163; c) Y. van den Winkel, H. Bastiaans, F. Bickelhaupt,
J. Organomet. Chem. 1991, 405, 183 – 194; d) M. Driess, Angew.
Chem. 1991, 103, 979 – 981; Angew. Chem. Int. Ed. Eng. 1991, 30,
1022 – 1024; e) M. Driess, H. Pritzkow, J. Chem. Soc. Chem.
Commun. 1993, 1585 – 1587.
! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
www.angewandte.org
!
!
Figure 3. Schematic representation of the bonding in HPSi in terms of
orbital interactions.
FT Microwave spectroscopy: The FT microwave spectrometer used
has been described in Ref. [20]. HPSi was formed in the throat of a
nozzle by applying a low-current dc discharge to a short gas pulse
created by a fast mechanical valve, the gas in the present work being a
mixture of phosphine (PH3) and silane (SiH4) heavily diluted (0.2 %)
in neon. To yield the strongest lines, the discharge potential was
1.2 kV, and the flow rate was about 20 cm3 min"1, at the 6 Hz pulse
rate of the nozzle with a stagnation pressure behind the valve of
2.5 kTorr.
Millimeter wave spectroscopy: The spectrometer used has been
described in Ref. [21]. HPSi was produced in a 1 m long glass cell,
from a mixture of elemental phosphorus, H2, and SiH4 diluted in
argon, in a 200 W ac glow discharge; D2 was used instead of H2 to
create DPSi. Typical pressures used were 20 mTorr of H2 and
20 mTorr of 0.8 % SiH4 in argon.
Quantum-chemical calculations: The calculations have been
performed using the CFOUR suite of programs.[22] The best-estimate
structure has been obtained at a level denoted as fc-CCSD(T)/ccwith
pV1Z + DT/cc-pVTZ + DQ/cc-pVDZ + Dcore/cc-pCV5Z[14]
the frozen-core CCSD(T) contribution extrapolated to the basis-set
limit (indicated by cc-pV1Z) and augmented by corrections for a full
CC singles, doubles, triples (CCSDT) treatment (using the cc-pVTZ
basis), for the effect of quadruple excitations at the CC singles,
doubles, triples, quadruples (CCSDTQ) level (using the cc-pVDZ
basis), and for core-valence correlation effects treated at the
CCSD(T)/cc-pCV5Z level. Vibrational corrections for the rotational
constants have been obtained at the CCSD(T)/cc-pwCVQZ level of
theory using second-order vibrational perturbation theory.[14] The
NBO analyses have been carried out at the CCSD/cc-pVTZ level
using the program package MOLPRO, version 2006.1.[23]
These are not the final page numbers!
Communications
[6] a) H. R. G. Bender, E. Niecke, M. Nieger, J. Am. Chem. Soc.
1993, 115, 3314 – 3315; b) M. Driess, S. Rell, H. Pritzkow, J.
Chem. Soc. Chem. Commun. 1995, 253 – 254.
[7] R. Varma, K. R. Ramaprasad, J. F. Nelson, J. Chem. Phys. 1975,
63, 915 – 918.
[8] a) M. Agfflndez, J. Cernicharo, J. R. Pardo, M. Gu#lin, T. G.
Phillips, Astron. Astrophys. 2008, 485, L33 – 36; b) E. D. Tenenbaum, L. M. Ziurys, Astrophys. J. 2008, 680, L121 – 124;
c) D. T. Halfen, D. J. Clouthier, L. M. Ziurys, Astrophys. J.
2008, 677, L101 – 104.
[9] a) A. D. McLean, J. Chem. Phys. 1962, 37, 627 – 630; b) J. B.
Ingels, J. M. Turney, N. A. Richardson, Y. Yamaguchi, H. F.
Schaefer III, J. Chem. Phys. 2006, 125, 104306; c) M. C. Lind,
F. C. Pickard IV, J. B. Ingels, A. Paul, Y. Yamaguchi, H. F.
Schaefer III, J. Chem. Phys. 2009, 130, 104301.
[10] a) K. J. Dykema, T. N. Troung, M. S. Gordon, J. Am. Chem. Soc.
1985, 107, 4535 – 4541; b) A. G. Baboul, H. B. Schlegel, J. Am.
Chem. Soc. 1996, 118, 8444 – 8451.
[11] C.-H. Lai, M.-D. Su, S.-Y. Chu, Inorg. Chem. 2002, 41, 1320 –
1322.
!
!
4
www.angewandte.org
[12] D. B. Chesnut, Chem. Phys. 2005, 315, 59 – 64.
[13] S. Thorwirth, J. Gauss, M. C. McCarthy, F. Shindo, P. Thaddeus,
Chem. Commun. 2008, 5292 – 5294.
[14] C. Puzzarini, M. Heckert, J. Gauss, J. Chem. Phys. 2008, 128,
194108.
[15] C. C. Loomis, M. W. P. Strandberg, Phys. Rev. 1951, 81, 798 – 807.
[16] K. Ohno, H. Matsuura, Y. Endo, E. Hirota, J. Mol. Spectrosc.
1986, 118, 1 – 17.
[17] A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88,
899 – 926.
[18] W. Kutzelnigg, Angew. Chem. 1984, 96, 262 – 286; Angew. Chem.
Int. Ed. Engl. 1984, 23, 272 – 295.
[19] T. Busch, W. W. Schoeller, E. Niecke, M. Nieger, H. Westermann, Inorg. Chem. 1989, 28, 4334 – 4340.
[20] M. C. McCarthy, W. Chen, M. J. Travers, P. Thaddeus, Astrophys.
J. Suppl. Ser. 2000, 129, 611 – 623.
[21] C. Savage, L. M. Ziurys, Rev. Sci. Instrum. 2005, 76, 043106.
[22] Additional information can be found at www.cfour.de.
[23] Additional information can be found at www.molpro.net.
! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers!
Angew. Chem. Int. Ed. 2010, 49, 1 – 5
Angewandte
Chemie
Communications
V. Lattanzi, S. Thorwirth, D. T. Halfen,
L. A. M"ck, L. M. Ziurys, P. Thaddeus,
J. Gauss,*
M. C. McCarthy*
&&&&—&&&&
Bonding in the Heavy Analogue of
Hydrogen Cyanide: The Curious Case of
Bridged HPSi
Angew. Chem. Int. Ed. 2010, 49, 1 – 5
Curiouser and curiouser: combining
high-resolution molecular spectroscopy
in the centimeter and millimeter wave
regions, and high-level coupled-cluster
quantum-chemical calculations, the
structure of the HPSi molecule has been
determined. The bridged geometry of
HPSi is in remarkable contrast to that of
the C and/or N analogues, such as HCN/
HNC, HCP, and HNSi which are all linear.
! 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5
www.angewandte.org
!
!
“Heavy” Hydrogen Cyanide
These are not the final page numbers!